Application of an Advanced Meta Model Selection Algorithm on the Sensitivity Analysis of a Cooled Turbine Blade

Abstract

Probabilistic methods are growing more important in the aerospace industry due to the ability to describe the behaviour of complex systems in the presence of input parameter variance. Sensitivity analysis based on meta models can be utilized for this purpose. The reliability of the results is dependent on the surrogate model quality, which in turn depends on the available data. A priori the appropriate meta model type is not known. An approach to automatically select the best fitting model for a given data set is presented in this paper. For comparison, polynomial regression with least squares fitting, moving least squares, radial basis functions, and support vector regression are used as candidate types. The selection of the best meta model type is based on two quality criteria utilizing a cross-validation (CV) scheme. The developed approach is demonstrated on the sensitivity analysis of a cooled turbine blade.